Anthropic is an artificial intelligence company based in San Francisco, California, and incorporated as a public-benefit corporation. It was founded in 2021 by a group of former OpenAI researchers and engineers, and it has grown into one of the most closely watched companies in the field. Its best known product is Claude, a family of large language models that people use through a chat interface and that developers reach through the Claude API.
The company describes its purpose as building reliable, interpretable, and steerable AI systems, with safety research treated as central rather than secondary to product work. That positioning has shaped how Anthropic recruits, what it publishes, and how it talks about the risks and benefits of increasingly capable models. The sections below cover what the company does, who founded it, the Claude product line, its safety approach, and the funding that has pushed its valuation into the hundreds of billions of dollars.
What Anthropic does
Anthropic researches and ships frontier AI models and sells access to them. Individuals interact with Claude through web and mobile apps and through Claude-powered features in other software, while companies and developers build on the Claude API to add language understanding, writing, analysis, and coding capability to their own products. The business mixes a consumer subscription, enterprise plans, and usage-based API pricing.
Alongside the commercial work, the company runs a research program that studies how large models behave and how to make them safer and more predictable. Anthropic publishes much of this research openly, covering topics such as interpretability, model evaluation, and methods for reducing harmful outputs. The two sides are meant to reinforce each other, with findings from safety research feeding back into the models that customers use.
Founding and team
Anthropic was founded in 2021 by siblings Dario Amodei and Daniela Amodei together with several colleagues who left OpenAI, including Jared Kaplan, Jack Clark, and Ben Mann. Dario Amodei, who had served as OpenAI's vice president of research, became chief executive officer, and Daniela Amodei became president. You can read more about the company's chief executive in our profile of Dario Amodei.
The founding group shared a view that the pace of AI progress warranted a dedicated effort focused on safety and reliability. From early on the company hired heavily from research backgrounds in machine learning, alignment, and policy, and it structured itself as a public-benefit corporation so that its stated mission carries legal weight alongside the interests of shareholders. That structure is part of how Anthropic explains the balance it tries to strike between commercial growth and its safety commitments.
Claude and products
Claude is Anthropic's family of large language models and the core of its product lineup. The models are offered at different sizes and capability levels so that users and developers can choose a balance of speed, cost, and reasoning ability. People reach Claude through the consumer Claude app, and developers integrate it through the Claude API and through major cloud platforms that resell access.
The product range has expanded beyond plain chat to include coding assistance, document analysis, and agentic features that let the model use tools and carry out multi-step tasks. Anthropic positions Claude for both everyday writing and research help and for demanding enterprise workloads, and it competes directly with offerings from OpenAI, Google, and other model developers. Pricing and model tiers are published on the company site so prospective users can compare options.
AI safety approach
Safety is the theme Anthropic returns to most often in describing itself. One of its signature contributions is Constitutional AI, a training method in which a model is guided by a written set of principles, or constitution, that it uses to critique and revise its own responses rather than relying solely on human feedback for every judgment. The aim is to make a model's behavior more transparent and easier to steer.
The company also invests in interpretability research, which tries to understand the internal workings of large models, and in evaluations that probe for dangerous capabilities before models are released. Anthropic has published a responsible scaling framework that ties safety measures to a model's capability level, committing to stronger safeguards as systems become more powerful. These practices are the basis for how the company argues it can build advanced AI while limiting the associated risks.
Funding and why it matters
Anthropic has raised large sums from venture investors, technology companies, and institutional funds, and its backers include major cloud and infrastructure partners. In its Series G round the company reported a $380 billion post-money valuation, and later private financing pushed the figure substantially higher as investors positioned ahead of a possible public listing. Those numbers place Anthropic among the most valuable private companies in the world.
The scale of this funding matters because frontier AI is capital intensive: training and serving large models requires enormous amounts of computing power, which in turn requires deep financing. For founders and operators watching the sector, Anthropic is a case study in how a safety-focused research lab built only a few years ago became a central commercial player. The company's trajectory toward the public markets is covered in our report on its confidential IPO filing.
Frequently asked questions
Who founded Anthropic and when?
Anthropic was founded in 2021 by a group of former OpenAI researchers, including siblings Dario Amodei and Daniela Amodei along with Jared Kaplan, Jack Clark, and Ben Mann. Dario Amodei serves as chief executive officer and Daniela Amodei as president.
What is Claude?
Claude is Anthropic's family of large language models. Individuals use it through the Claude app, and developers build on it through the Claude API and major cloud platforms.
Where is Anthropic headquartered?
Anthropic is headquartered in San Francisco, California, in the United States.
What is Constitutional AI?
Constitutional AI is a training method Anthropic developed in which a model follows a written set of principles, or constitution, to critique and revise its own responses, which is intended to make its behavior more transparent and easier to steer.
How much is Anthropic worth?
Anthropic reported a $380 billion post-money valuation in its Series G round, and later private financing valued the company considerably higher as it moved toward a possible public listing. Figures change with each new round, so the most current valuation is best checked against recent reporting.
Sources and references
Company information and product details are published on the official site at anthropic.com. Anthropic announced its Series G round and $380 billion valuation in its own funding announcement. Background on the company's founding and history is summarized in Britannica's profile of Anthropic.